CN104331712A - Automatic classifying method for algae cell images - Google Patents
Automatic classifying method for algae cell images Download PDFInfo
- Publication number
- CN104331712A CN104331712A CN201410682794.4A CN201410682794A CN104331712A CN 104331712 A CN104331712 A CN 104331712A CN 201410682794 A CN201410682794 A CN 201410682794A CN 104331712 A CN104331712 A CN 104331712A
- Authority
- CN
- China
- Prior art keywords
- algae
- class
- represent
- classification
- image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
- G06F18/2155—Generating training patterns; Bootstrap methods, e.g. bagging or boosting characterised by the incorporation of unlabelled data, e.g. multiple instance learning [MIL], semi-supervised techniques using expectation-maximisation [EM] or naïve labelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/217—Validation; Performance evaluation; Active pattern learning techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses an automatic classifying method for algae cell images. The automatic classifying method includes steps that pre-processing algae images to obtain processed algae images; extracting visual features from the algae images according to the processed algae images to obtain feature vectors expressed by multi-feature fusion; classifying all the algae cells according to category, extracting the cell image with special proportion from each type of cells, and using the extracted cell images as a subsequent model learning training set; constructing target functions of a multi-task learning mode and solving; using a model obtained through training to forecast test data. The automatic classifying method for algae cell images lowers the calculating complexity and improves the algae cell classifying rate.
Description
Technical field
The present invention relates to the fields such as image procossing, pattern classification, environmental monitoring, particularly relate to a kind of alga cells classification of images method.
Background technology
Along with the development of computer science and the demand of cross discipline application, computing technique has been widely used in environmental monitoring field.Deepening continuously particularly along with artificial intelligence field research, to understand automatically the media content such as image, video become possibility by realizing computing machine to the structure of specific perpetual object mathematical model.Therefore, the subject crossing research between environmental science and information science can promote advanced environment detection method and the research of instrument.At present, the classification of Measures of Algae in Water Body cell is monitoring water pollutant and the important references index differentiating Water quality with counting.But, only undertaken classifying and counting by manpower in current environment monitoring, the method wastes time and energy, and due to many factors such as experiences, current this height depends on subjectivity and sentences the accurate identification that method for distinguishing often can not realize alga cells classification fast and accurately, and therefore realizing alga cells classification of images by the method for image procossing and pattern-recognition becomes problem demanding prompt solution.
Less about the method for algae image automatic classification at present, according to the difference adopting model, existing method is divided into the following two kinds:
1) based on the method [1-3] of static model: the method often extracts the visual signature (as: feature such as color, texture and shape) in specific cells region, then utilize common static classifiers (as support vector machine, naive Bayesian etc.) carry out model learning, and then automatic classification is carried out to the algae image of unknown classification.But these class methods often depend on the visual signature descriptor with robust resolution characteristic, otherwise be difficult to the alga cells of Accurate classification deformation complexity.
2) based on the method [4] of temporal model: these class methods often extract the characteristic sequence characterizing the continuous moment metamorphosis of certain class cell, then the graph model of characterization time sequence is passed through (as hidden Markov model, random field models etc.) carry out sequence dynamic learning, then utilize and train the temporal model obtained to carry out cell classification.But the method carries out having mass data when temporal model learns relying on very by force, and calculated amount is large because model structure complexity causes study, and speed is slow.
Summary of the invention
The invention provides a kind of alga cells classification of images method, present invention reduces computation complexity, improve the classification rate of alga cells, described below:
A kind of alga cells classification of images method, described automatic classification method comprises the following steps:
Pre-service is carried out to algae image, obtains the algae image after process;
According to the algae image after process, carry out algae image Visual Feature Retrieval Process, obtain the proper vector that multiple features fusion represents;
All alga cells are classified according to classification, the training set that the cell image extracting the special ratios in every class cell learns as following model;
Build the objective function of multi-task learning pattern and solve; Use and train the model obtained to predict test data;
Wherein, objective function is:
In formula,
represent the set of M task model parameter, W
jfor a jth column vector of W, represent a jth model parameter,
represent the multitask combination learning part based on least square,
represent the regular terms based on the sparse optimization thought design of group; || ||
2represent the l of column vector
2norm, η is weight parameter,
represent the feature of a kth sample of jth class cell, Y
jrepresent jth class class label.
Described use trains the model obtained to be specially the step that test data is predicted:
Adopt the model W of l class
lto sample F to be tested
ucalculate W
l× F
umark with all categories
the absolute value of error | W
l× F
u-Y
j|, record now least error E
l, after calculating least error corresponding to all class models successively, select the prime marks p that minimum value in these values is corresponding, by F
ube categorized as p class; If Y
j=p, then correctly classify, otherwise be mis-classification.
The beneficial effect of technical scheme provided by the invention is: all alga cells are classified according to classification by the present invention, the training set that the cell image extracting the special ratios in every class cell learns as following model; Build the objective function of multi-task learning pattern and solve; Use and train the model obtained predict test data thus solve the dependence of first kind method to high resolution visual signature, Equations of The Second Kind method can be avoided again the dependence of large data and high computation complexity.Present invention reduces computation complexity, improve the classification rate of alga cells.
Accompanying drawing explanation
Fig. 1 is the schematic diagram of rhombus algae;
Fig. 2 is the schematic diagram of the curved algae of bridge;
Fig. 3 is the schematic diagram of two eyebrow algae;
Fig. 4 is a kind of process flow diagram of alga cells classification of images method.
Embodiment
For making the object, technical solutions and advantages of the present invention clearly, below embodiment of the present invention is described further in detail.
This method independently carries out the existing method of modeling to every class algae in background technology by breaking through, adopt the model of multitask combination learning to carry out model learning to multiple algae simultaneously, thus solve first kind method to the dependence of high resolution visual signature, Equations of The Second Kind method can be avoided again the dependence of large data and high computation complexity.
101: pre-service is carried out to algae image, obtain the algae image after process;
Wherein, to inputted algae image, (only comprise an alga cells in piece image, the type of the parameters such as harvester as shown in Figure 1), camera parameter, image storage format and image resolution ratio and algae image is all without particular restriction.
In order to meet the data mode needed for subsequent treatment, the algae image of this method to input carries out image gray processing process, then to then carrying out filtering to image after gray proces, its objective is and filter picture noise, filtered graphical rule is normalized, realize for bilinear interpolation in the present invention, specifically do not limit.
The above-mentioned algae image to input carries out image gray processing process successively, the step of filtering process and normalized is conventionally known to one of skill in the art, and the embodiment of the present invention does not limit this.Such as: the image gray processing process in this method adopts the content in document [5], filtering process and normalized adopt the content in document [6].
102: according to the algae image after process, carry out algae image Visual Feature Retrieval Process;
Because the method for follow-up employing multitask combination learning improves model resolving ability, so the present invention does not do concrete restriction for algae image visual signature classification, the texture embodied for algae image in upper figure in experiment and shape facility, have employed multiple features fusion strategy.
That is: to image I, extract N class visual signature respectively, the i-th class visual signature adopts vectorial Fi to represent, is then cascaded up by N category feature vector, formed proper vector F=that final multiple features fusion represents [F1, F2 ..., FN]
Because histograms of oriented gradients has very strong performance for sign alga cells outer contour shape, so be extracted histograms of oriented gradients proper vector F1 to algae image I, specifically can content in list of references [7], the present invention does not repeat this.
Ability is characterized very by force because Scale invariant features transform SIFT has for the local grain characteristic of image, adopt Scale invariant features transform SIFT special to vectorial F2 so be extracted algae image I, specifically can content in list of references [8], the present invention does not repeat this.
On this basis, the two is cascaded up, form the proper vector [F1, F2] after merging.
103: all alga cells are classified according to classification, the training set that the cell image extracting the special ratios in every class cell learns as following model;
All alga cells are classified according to classification, supposes M class alga cells altogether.Jth class cell Sj represents, wherein
represent the character representation of a kth sample of jth class cell, wherein such cell has N
jindividual sample.Every class cell formulates identical class label Y
j, all categories label sets is designated as Y.
Then, the training set that the cell image extracting the special ratios R in every class cell learns as following model, all the other all data are used as test data set.Wherein, ratio R sets as required, and in experiment, reference value is 50%.The method of data decimation also can be selected according to actual conditions, and in experiment, reference method is artificial Stochastic choice.
104: build the objective function of multi-task learning pattern and solve;
At present, existing sorter major part is all identify some algae separately, if the identification of each algae is regarded as a task, this sorting technique is called single tasking learning.In single tasking learning process, each task is considered to independently to carry out, and have ignored the relevance between task.Therefore, this method adds the related information between algae, carries out Classification and Identification, i.e. multi-task learning to multiple algae simultaneously.In multi-task learning process, multiple inter-related task learns simultaneously, realizes the information sharing between multitask, indirectly increases the number of samples of participation task, improves the performance of prediction.Therefore, multi-task learning is highly profitable, especially under the situation that the training sample of database is little to the accuracy rate improving alga classifying.
The Least-squares minimization form of each task model added up in the present invention, the multitask combination learning method formed based on least square carries out the training of multi task model.In addition, in order to excavate the associate feature between multiple task, this method has used for reference group sparse optimization thought realizes every generic task model parameter selection by the design of regular terms, thus indirectly realize the selection of characteristics of image, the disadvantageous feature of association mining will be combined to multitask to filter, thus be more conducive to the model of different algal species automatic classification.
The method establishing target function is:
Wherein,
represent the set of M task model parameter, wherein W
jfor a jth column vector of W, represent a jth model parameter.Here,
represent the multitask combination learning part based on least square,
represent the regular terms based on the sparse optimization thought design of group.Wherein, || ||
2represent the l of column vector
2norm.η is weight parameter, usually carries out empirical setting.Empirically in experiment have employed 0.01.
This multi-task learning model can solve by combining the objective function optimization of above-mentioned least square method pattern, and specific algorithm does not limit.In order to solve above-mentioned convex optimization object function in experiment, have employed famous convex optimization tool bag CVX and solving, see list of references [9].
105: use and train the model obtained to predict test data.
In the present invention, propose the multitask sorting technique based on global optimum.The present invention adopts the model W of l class successively
lto sample F to be tested
u(sample to be tested selected from test set supposes during test that this sample class is unknown, so use F
urepresent) calculate W
l× F
umark with all categories
the absolute value of error | W
l× F
u-Y
j|, record now least error E
l(representing error when selection l class model), after calculating least error corresponding to all class models successively, select the prime marks p that minimum value in these values is corresponding, can by F
ube categorized as p class.
Adopt said method to carry out class prediction to all samples in each class testing data, then add up the accuracy rate of this alga cells classification, concrete grammar is as follows:
With jth alga cells
be predicted as example:
To the sample of the Nj in jth alga cells respectively with said method prediction, the prediction classification of each sample is designated as p, and time in step 103, sample prepares, its true class label is Y
jif, Y
j=p, then correctly classify, otherwise be mis-classification.Add up correct classification samples sum N
tP, then jth class cell classification accuracy rate is N
tP/ N
j.
All alga cells classification accuracies can be counted on by above-mentioned steps 101-105.
The feasibility of this method is verified below by concrete several groups experiments, described below:
In experiment, prepared 6 class algae image by cultivating, be respectively: woven design algae, rhombus algae, two eyebrow algae, two water chestnut algae, melosira, the curved algae of bridge, every class has prepared 1000 samples, and image resolution ratio is 100 × 100.The 3 kinds of methods choosing comparative maturity in prior art are tested as a comparison, and classified to 6 class algae image by above-mentioned steps 101-105, the classification results obtained is as shown in table 1:
Table 1
Wherein, control methods 1,2 and 3 is respectively the method described in list of references [10], [11] and [12], and those methods are conventionally known to one of skill in the art.Can find out that the classify accuracy of this method is apparently higher than other three kinds of methods of the prior art, demonstrates the feasibility of this method, can meet the needs in practical application by above-mentioned experiment.
List of references
1.Liu,A.,Li,K.,Kanade,T.:Spatiotemporal mitosis event detection in time-lapse phase contrast microscopy image sequences.In:IEEE International Conference on Multimedia and Expo,pp.1–6(2010)
2.Perner,P.,Perner,H.,Mller,B.:Mining knowledge for hep-2 cell image classification.Artificial Intelligence in Medicine 26,161–173(2002)
3.Cordelli,E.,P.,S.:Color to grayscale staining pattern representation in iif.In:In International Symposium on Computer-Based Medical Systems,pp.1–6(2011)
4.Huh,S.,Ker,D.F.E.,Bise,R.,Chen,M.,Kanade,T.:Automated mitosis detection of stem cell populations in phase-contrast microscopy images.IEEE Transactions on Medical Imaging pp(99),1–12(2010)
5. Ruan Qiu fine jade, Ruan Yuzhi, Digital Image Processing (the 3rd edition), on May 1st, 2011.
6. a moral is rich etc. writes, MATLAB Digital Image Processing, China Machine Press, 2009.
7.Dalal N,Triggs B.Histograms of oriented gradients for human detection[C].IEEE International Conference on Computer Vision and Pattern Recognition,886–893,2005
8.Lowe,D.G.:Distinctive image features from scale-invariant keypoints.Int.J.Computer Vision 60(2),91–110,(2004)
9.Michael Grant and Stephen Boyd.CVX:Matlab software for disciplined convex programming,version 2.0 beta.http://cvxr.com/cvx,September 2013.
10.Liu,A.,Li,K.,Kanade,T.:Spatiotemporal mitosis event detection in time-lapse phase contrast microscopy image sequences.In:IEEE International Conference on Multimedia and Expo,pp.1–6(2010)
11.Perner,P.,Perner,H.,Mller,B.:Mining knowledge for hep-2 cell image classification.Artificial Intelligence in Medicine 26,161–173(2002)
12.Cordelli,E.,P.,S.:Color to grayscale staining pattern representation in iif.In:In International Symposium on Computer-Based Medical Systems,pp.1–6(2011)
It will be appreciated by those skilled in the art that accompanying drawing is the schematic diagram of a preferred embodiment, the invention described above embodiment sequence number, just to describing, does not represent the quality of embodiment.
The foregoing is only preferred embodiment of the present invention, not in order to limit the present invention, within the spirit and principles in the present invention all, any amendment done, equivalent replacement, improvement etc., all should be included within protection scope of the present invention.
Claims (2)
1. an alga cells classification of images method, is characterized in that, described automatic classification method comprises the following steps:
Pre-service is carried out to algae image, obtains the algae image after process;
According to the algae image after process, carry out algae image Visual Feature Retrieval Process, obtain the proper vector that multiple features fusion represents;
All alga cells are classified according to classification, the training set that the cell image extracting the special ratios in every class cell learns as following model;
Build the objective function of multi-task learning pattern and solve; Use and train the model obtained to predict test data;
Wherein, objective function is:
In formula,
represent the set of M task model parameter, W
jfor a jth column vector of W, represent a jth model parameter,
represent the multitask combination learning part based on least square,
represent the regular terms based on the sparse optimization thought design of group; || ||
2represent the l of column vector
2norm, η is weight parameter,
represent the character representation of a kth sample of jth class cell, Y
jrepresent jth class class label.
2. a kind of alga cells classification of images method according to claim 1, is characterized in that, described use trains the model obtained to be specially the step that test data is predicted:
Adopt the model W of l class
lto sample F to be tested
ucalculate W
l× F
umark with all categories
the absolute value of error | W
l× F
u-Y
j|, record now least error E
l, after calculating least error corresponding to all class models successively, select the prime marks p that minimum value in these values is corresponding, by F
ube categorized as p class; If Y
j=p, then correctly classify, otherwise be mis-classification.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410682794.4A CN104331712B (en) | 2014-11-24 | 2014-11-24 | A kind of alga cells classification of images method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410682794.4A CN104331712B (en) | 2014-11-24 | 2014-11-24 | A kind of alga cells classification of images method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104331712A true CN104331712A (en) | 2015-02-04 |
CN104331712B CN104331712B (en) | 2017-08-25 |
Family
ID=52406433
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410682794.4A Expired - Fee Related CN104331712B (en) | 2014-11-24 | 2014-11-24 | A kind of alga cells classification of images method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104331712B (en) |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106022303A (en) * | 2016-06-06 | 2016-10-12 | 南昌航空大学 | Fresh water algae rough classification and counting method based on lensless holographic imaging |
CN106446927A (en) * | 2016-07-07 | 2017-02-22 | 浙江大学 | Self-paced reinforcement image classification method and system |
CN109271977A (en) * | 2018-11-23 | 2019-01-25 | 四川长虹电器股份有限公司 | The automatic classification based training method, apparatus of bill and automatic classification method, device |
CN109447950A (en) * | 2018-09-30 | 2019-03-08 | 北京英视睿达科技有限公司 | A kind of method of counting and system for microorganism |
CN109858386A (en) * | 2019-01-09 | 2019-06-07 | 深圳市趣方科技有限公司 | A kind of microalgae cell recognition methods based on fluorescence microscope images |
CN110532941A (en) * | 2019-08-27 | 2019-12-03 | 安徽生物工程学校 | A kind of characteristic image extracting method of common algae |
WO2020042704A1 (en) * | 2018-08-27 | 2020-03-05 | 杭州德适生物科技有限公司 | Chromosome identifying method based on deep learning |
JPWO2018207885A1 (en) * | 2017-05-10 | 2020-03-12 | 日本電気株式会社 | Analysis device, stratum age estimation device, analysis method, stratum age estimation method, and program |
CN111368872A (en) * | 2019-12-20 | 2020-07-03 | 浙江大学 | Breast cancer mitosis cell detection method based on fusion characteristics and verification model |
CN111524138A (en) * | 2020-07-06 | 2020-08-11 | 湖南国科智瞳科技有限公司 | Microscopic image cell identification method and device based on multitask learning |
TWI715877B (en) * | 2017-10-10 | 2021-01-11 | 大陸商杭州智微信息科技有限公司 | Method and system for marking bone marrow cells |
CN113268497A (en) * | 2020-12-15 | 2021-08-17 | 龚文凯 | Intelligent recognition learning training method and device for key target parts |
CN115925076A (en) * | 2023-03-09 | 2023-04-07 | 湖南大学 | Coagulation automatic dosing method and system based on machine vision and deep learning |
CN117274869A (en) * | 2023-09-25 | 2023-12-22 | 北方工业大学 | Cell deformation dynamic classification method and system based on deformation field extraction |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102663406A (en) * | 2012-04-12 | 2012-09-12 | 中国海洋大学 | Automatic chaetoceros and non-chaetoceros sorting method based on microscopic images |
CN102682302A (en) * | 2012-03-12 | 2012-09-19 | 浙江工业大学 | Human body posture identification method based on multi-characteristic fusion of key frame |
US20120251008A1 (en) * | 2011-03-31 | 2012-10-04 | Microsoft Corporation | Classification Algorithm Optimization |
CN102722717A (en) * | 2012-05-22 | 2012-10-10 | 天津大学 | Cell division identification method |
CN103605667A (en) * | 2013-10-28 | 2014-02-26 | 中国计量学院 | Automatic image annotation algorithm |
-
2014
- 2014-11-24 CN CN201410682794.4A patent/CN104331712B/en not_active Expired - Fee Related
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20120251008A1 (en) * | 2011-03-31 | 2012-10-04 | Microsoft Corporation | Classification Algorithm Optimization |
CN102682302A (en) * | 2012-03-12 | 2012-09-19 | 浙江工业大学 | Human body posture identification method based on multi-characteristic fusion of key frame |
CN102663406A (en) * | 2012-04-12 | 2012-09-12 | 中国海洋大学 | Automatic chaetoceros and non-chaetoceros sorting method based on microscopic images |
CN102722717A (en) * | 2012-05-22 | 2012-10-10 | 天津大学 | Cell division identification method |
CN103605667A (en) * | 2013-10-28 | 2014-02-26 | 中国计量学院 | Automatic image annotation algorithm |
Cited By (21)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106022303A (en) * | 2016-06-06 | 2016-10-12 | 南昌航空大学 | Fresh water algae rough classification and counting method based on lensless holographic imaging |
CN106022303B (en) * | 2016-06-06 | 2019-07-19 | 南昌航空大学 | Fresh water algae rough sort and method of counting based on no lens holographic imaging |
CN106446927A (en) * | 2016-07-07 | 2017-02-22 | 浙江大学 | Self-paced reinforcement image classification method and system |
CN106446927B (en) * | 2016-07-07 | 2019-05-28 | 浙江大学 | It is a kind of to enhance image classification method and system from step |
JPWO2018207885A1 (en) * | 2017-05-10 | 2020-03-12 | 日本電気株式会社 | Analysis device, stratum age estimation device, analysis method, stratum age estimation method, and program |
US11448634B2 (en) | 2017-05-10 | 2022-09-20 | Nec Corporation | Analysis apparatus, stratum age estimation apparatus, analysis method, stratum age estimation method, and program |
TWI715877B (en) * | 2017-10-10 | 2021-01-11 | 大陸商杭州智微信息科技有限公司 | Method and system for marking bone marrow cells |
WO2020042704A1 (en) * | 2018-08-27 | 2020-03-05 | 杭州德适生物科技有限公司 | Chromosome identifying method based on deep learning |
US11436493B2 (en) | 2018-08-27 | 2022-09-06 | Hangzhou Diagens Biotech Co., LTD. | Chromosome recognition method based on deep learning |
CN109447950A (en) * | 2018-09-30 | 2019-03-08 | 北京英视睿达科技有限公司 | A kind of method of counting and system for microorganism |
CN109447950B (en) * | 2018-09-30 | 2023-12-15 | 国际商业机器(中国)投资有限公司 | Counting method and system for microorganisms |
CN109271977A (en) * | 2018-11-23 | 2019-01-25 | 四川长虹电器股份有限公司 | The automatic classification based training method, apparatus of bill and automatic classification method, device |
CN109858386A (en) * | 2019-01-09 | 2019-06-07 | 深圳市趣方科技有限公司 | A kind of microalgae cell recognition methods based on fluorescence microscope images |
CN110532941A (en) * | 2019-08-27 | 2019-12-03 | 安徽生物工程学校 | A kind of characteristic image extracting method of common algae |
CN111368872A (en) * | 2019-12-20 | 2020-07-03 | 浙江大学 | Breast cancer mitosis cell detection method based on fusion characteristics and verification model |
CN111524138A (en) * | 2020-07-06 | 2020-08-11 | 湖南国科智瞳科技有限公司 | Microscopic image cell identification method and device based on multitask learning |
CN113268497A (en) * | 2020-12-15 | 2021-08-17 | 龚文凯 | Intelligent recognition learning training method and device for key target parts |
CN115925076A (en) * | 2023-03-09 | 2023-04-07 | 湖南大学 | Coagulation automatic dosing method and system based on machine vision and deep learning |
CN115925076B (en) * | 2023-03-09 | 2023-05-23 | 湖南大学 | Automatic coagulation dosing method and system based on machine vision and deep learning |
CN117274869A (en) * | 2023-09-25 | 2023-12-22 | 北方工业大学 | Cell deformation dynamic classification method and system based on deformation field extraction |
CN117274869B (en) * | 2023-09-25 | 2024-03-26 | 北方工业大学 | Cell deformation dynamic classification method and system based on deformation field extraction |
Also Published As
Publication number | Publication date |
---|---|
CN104331712B (en) | 2017-08-25 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104331712A (en) | Automatic classifying method for algae cell images | |
He et al. | An end-to-end steel surface defect detection approach via fusing multiple hierarchical features | |
CN109952614B (en) | Biological particle classification system and method | |
EP4027300A1 (en) | Apparatus, program, and method for anomaly detection and classification | |
Katharopoulos et al. | Processing megapixel images with deep attention-sampling models | |
CN111126386B (en) | Sequence domain adaptation method based on countermeasure learning in scene text recognition | |
Colak et al. | Automated McIntosh-based classification of sunspot groups using MDI images | |
CN101944174B (en) | Identification method of characters of licence plate | |
CN104834941A (en) | Offline handwriting recognition method of sparse autoencoder based on computer input | |
Wang et al. | A computer vision based machine learning approach for fatigue crack initiation sites recognition | |
Zhang et al. | Self-guided adaptation: Progressive representation alignment for domain adaptive object detection | |
CN105930792A (en) | Human action classification method based on video local feature dictionary | |
Cao et al. | Detection of microalgae objects based on the Improved YOLOv3 model | |
Alhroob et al. | Fuzzy min-max classifier based on new membership function for pattern classification: a conceptual solution | |
CN104598898A (en) | Aerially photographed image quick recognizing system and aerially photographed image quick recognizing method based on multi-task topology learning | |
Eerola et al. | Survey of automatic plankton image recognition: challenges, existing solutions and future perspectives | |
CN104537383A (en) | Massive organizational structure data classification method and system based on particle swarm | |
Rethik et al. | Attention Based Mapping for Plants Leaf to Classify Diseases using Vision Transformer | |
CN104537392A (en) | Object detection method based on distinguishing semantic component learning | |
Wang et al. | Recyclable waste identification using cnn image recognition and gaussian clustering | |
Rakowski et al. | Hand shape recognition using very deep convolutional neural networks | |
Ridoy et al. | A lightweight convolutional neural network for white blood cells classification | |
Shishkin et al. | Implementation of yolov5 for detection and classification of microplastics and microorganisms in marine environment | |
Bruch et al. | Evaluation of semi-supervised learning using sparse labeling to segment cell nuclei | |
MAIYANTI et al. | ROTATION-GAMMA CORRECTION AUGMENTATION ON CNN-DENSE BLOCK FOR SOIL IMAGE CLASSIFICATION |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170825 Termination date: 20191124 |